Meta-Analysis of 40 Machine Learning Models for Aortic Dissection Risk

40 studies on machine learning models for aortic dissection risk prediction were systematically reviewed and meta-analyzed. The analysis, conducted by Yijun Mao, synthesized performance metrics like pooled C-statistics for outcomes including early mortality and acute kidney injury.

Use Cases

• Compare pooled C-statistics (AUC) across outcomes like early mortality and long-term mortality from the meta-analysis
• Analyze the distribution of machine learning algorithms (e.g., random forest, XGBoost) used across the 40 included studies
• Assess model reporting quality against TRIPOD standards using extracted adherence percentages
• Investigate the prevalence of high/unclear risk of bias as identified by the PROBAST assessment across all studies

Strengths

• Meta-analysis synthesizes data from 40 distinct studies on a critical medical topic
• Performance metrics are provided for specific outcomes, such as a pooled C-statistic of 0.891 for early mortality
• Adherence to systematic review protocols (PRISMA, CHARMS) is documented in the source description

Limitations

• The dataset is a summary table (16.5 KB) with no raw patient-level data for model training
• The source review notes significant heterogeneity (I² = 61.3–78.8%) and high risk of bias in all included studies
• Critical model specifications and predictor definitions were often unreported in the original studies

Provenance

Source

Systematic review and meta-analysis by Yijun Mao, published on figshare.

Collection Method

Literature search across six databases (PubMed, Web of Science, etc.) from inception to September 2025, following PRISMA/CHARMS guidelines.

Time Range

Studies from database inception to September 30, 2025.

Freshness

Last updated March 2026.

Geography

null